Bitcoin Forum

I still see 2140 given as the date for new bitcoins to run out, in all kinds of places.

Why is this date so prevalent and so stuck in people's heads?

Do people not understand the way difficulty readjustments work? Every difficulty increase happens because 2016 blocks were found in shorter time than projected. The timeline is based on 10 minute per block projection. Every difficulty increase moves forward the date when all new bitcoins have been minted, unless we have some future difficulty decreases that compensate for the difficulty increase.

So why do we keep saying 2140? Closer to 2100 is more accurate, IMHO.

Someone made a study that it's closer to 2050.

come the next generation of asics, it might be within the next 3-10 years imo. i'm curious to know what happens to the value once mining has ended. does it die because of lack of miners, or does the value increase?

If >90% happens once within a block adjustment, difficulty will ratchet up, and move the end date forward 12 days. If it doesn't happen again the next adjustment, things will settle out and we're back on the slow march. I see 2050 as possible if we have constant ramp-up in hash power, but not 3-10 years. For it to be 3-10 years we would have to have CONSTANT increases in hash power of >90%. I just don't see that happening.

If we move forward 2 days every difficulty adjustment (average difficulty adjustment of ~15%) that moves the end date forward by about 18 years. 3 days each adjustment (average difficulty increase of ~21%) would move the end date forward by about 27 years.

2140 is right!

Oh, I know. I've seen the time-tables. My question is why do we keep insisting on that?

Is it because the unknown variable of new coins running out weighs heavy on everyone, so we keep the date of its occurrence as far away as conceivable?

This discussion is irrelevant.

What counts is at which point in time there are almost no new coins to be injected. That is, at which point is the influx of new coins no longer an economic factor. Unfortunately this is highly dependant on the development of the rates and the fraction of coins in circulation and thus pure speculation. And this explains why everyone concentrates on the irrelevant, yet more predictable end-of-new-coin date.

I actually agree with you that the end of coin date is irrelevant. The point where transaction fees consistently exceed new coins in the block reward is much more important, and much harder to predict.

And I think you have a point that the difficulty of predicting that point is why people talk about the end of new coins date instead, but it still doesn't explain why we keep saying 2140 when everyone who understands difficulty adjustment knows it will be sooner than that.

The Mayans predicted all the Bitcoins would be produced on December 21, 2012

+1

fail.
based on a 10 minute timeframe the total 21mill bitcoins will be found beyond 2200, but the dates at which the block reward decreases to such a small amount of satoshi's per block to even be worthy of mining, will come sooner

but before we get to that lets address the quoted fail..

lets say based on todays difficulty. a number of ASICS far surpass the estimate of the last difficulty increase and so blocks are now found in 9 minutes or less.. the next change will push the difficulty to a 12 minute rate per block solution to ensure that the solving of blocks stay on their 4 year half life cycle.
and lets say there was a period where miners never exceeded the estimate. then the difficulty would not increase. it would stay the same or decrease, ensuring that at the end of the 4 year the allotment of coins for that 4 year period were rewarded.

what you will find is that in 2037 each 10 minute reward would be worth 1 satoshi meaning unless extra decimals are added, the end would be 2141 for reward mining. even though there is still only 20,999,999 btc in existence (not quite finished).

alot of miners will think its time to end mining in 2029-2033 when the reward moves from just over 1.5btc to 0.7812500 per 10 minutes but thats because they are basing the profitability etc on todays value and hypertheticals

Bullshit.

They did NOT. They don't even known about SHA256 algorithms.

What this guy said.

also
Lol :D

Maybe not. You can buy more with the 25 reward today than you could with the 50 reward at the start.

ding ding ding we a have a winner. Glad a few people get it :)

I don't care when the last satoshi will be mined. What really matters is this:

90% of all coins will be minted by 2023 (in ten years).
98% of all coins will be minted by 2033 (in twenty years).

In 2023 the annual inflation rate will be around 2-3%.
In 2033 the annual inflation rate will be less than 1%.

So in just 10-20 years the inflation will be so insignificant, newly minted coins won't make any difference on the market. If fiat money happen to survive for the next 20 years, miners won't be able to produce 10-20% price volatility due to selling BTC to pay for equipment.

Illustration by Matt Whitlock:
http://www.mattwhitlock.com/Bitcoin%20Inflation.png

damn I'm kinda late :/

Nothing grows at 10%+ every 2 weeks for a century.  Run the numbers out for a decade or two if you don't believe me.

http://imgs.xkcd.com/comics/extrapolating.png

This is false. From the wiki:

Following this explanation, time should be 1209600 seconds. if time is actually 1000000 seconds, retarget would be 1209600/1000000 = 1.2096 giving new difficulty of current_difficulty*1.2096 = new_difficulty.
That would project forward 1209600 seconds for the next 2016 blocks, based on the average finding time of the previous 2016 blocks.

If this is wrong, show me code-blocks, not extrapolation.

As for the 10% every 2 weeks for a century... agreed. Nothing grows at that rate forever. Mean reversion suggests we will stabilize between 5-10% annually at some point. The moving forward of the date which has already occurred, though, is unlikely to ever be reversed.

definetly a winner hense the

So what this potentially means that if bitcoins keep the newcomer demand high, and the supply falls, there will be a great rise in the bitcoin price on a daily basis.  And in just 10 years from now, 1 bitcoin could be worth 10k or more (a bit optimistic).

If bitcoin demand grows, the price will grow. If it doesn't (market is saturated), then price stops growing.

With more than 50% coins already mined and quite low annual inflation (12% in 2013 comparing to 50% in 2011), supply does not affect prices much (in 2011 if you remember the price was slowly going down as miners were steadily selling their coins). Today demand is defining the price much more.

If we are doing optimistic estimation, it depends on the target market. If Bitcoin takes half of gold market as a distributed ledger of wealth, then its price has to be more than 10-20K USD. If it takes 10% of the global black market (estimated size is 1800 bln USD), its price must be around 8000 USD per BTC. Choose your market and figure how much BTC must be worth if it's used on X% of that market. If BTC replaces all currencies on global scale, each BTC would be much more than 1 million of today's dollars. I believe this can happen, but it may take a generation or two.

Can anyone confirm if this is true? It sounds wrong.

It is wrong. Next difficulty is adjusted based on mean speed of the previous 2016 blocks. When hashpower monotonically grows, it means that blocks will still appear quicker than 10 minutes, but not that much quicker than before the adjustment.

Notice how difficulty is always behind the growing hashpower.

http://bitcoin.sipa.be/speed-small-lin-10k.png

PS. Those who think 10 min interval is sacred and typical intervals of 8 minutes are "evil" do not understand anything and should go and re-read bitcoin.org/bitcoin.pdf multiple times. 10 min interval is a good enough approximation that makes miners waste less than 1% of CPU time while newly mined block is being propagated. 10 min is also good for sending blocks over partisan radio networks or to the satellite (or a google baloon). In itself, 10 or 8 minutes per block it does not matter at all. When the inflation rate drops to small numbers (<5%), it won't matter how fast blocks are actually appearing: every 10 minutes, 7 or 5.

I also responded to that at the bottom of page 1. I think it got missed somehow.

I like the chart, though. It's a very good illustration of the points I was trying to make.

my use of the 9 minutes -12 minutes were not to be taken literally. they were used to indicate that the difficulty would change to get an average time of 10 minutes per 4032 block.

hmmm..

my use of the amount 10 minutes and 4032 blocks is not to be taken literally, it is used to indicate that the difficulty over a average time evens out to ensure that the block solving stays on course for its 4 year cycle before the reward halving.

there is infact a slight variance on the number of seconds that has been mathemaically calculated into the bitcoin rules to make it not quite 10 minutes.

(hopefully this reply should quash the anal dwelling but monkeys that want to take things literally, instead of reading and putting into context)